We demonstrated that when a substantial fraction of a population is genotyped with a high-density SNP array, there is much more information in the data than what lies on the surface. Compared to existing approaches that either focus on close relatives, i.e. family-based analyses, or very distantly related (‘unrelated’) individuals, i.e. LD-based analyses, the conceptual leap here lies in the recognition of the utility of moderately to distantly related individuals, e.g. those separated by 3 to 20 or more meioses. As the number of meioses increases, the decrease in the probability of IBD sharing for any particular relative is compensated for by the exponential increase in the number of such relatives. While it is generally recognized among investigators familiar with linkage analysis that IBD sharing can often be detected without precise knowledge of the pedigree structure, it still came as a surprise when we realized that the information could be exploited in this systematic and extensive manner.
